search

asrhou / NATMI

NATMI: Network Analysis Toolkit for the Multicellular Interactions
MIT License
16 stars 11 forks source link

Calculation of "Log2-transformed fold change of edge expression weight" and non-differentially expressed interacting partners. #29

Open charlesgwellem opened 4 years ago

charlesgwellem commented 4 years ago

Hi. I have used NATMI on a single cell RNA sequencing dataset to perform a differential cell communication analysis between a data set in two conditions and one of the files that is output is "DOWN-regulated_mean.csv". After noticing how high the log2 -transformed values were, I think I do not understand how the "Log2-transformed fold change of edge expression weight" column is calculated. I try to calculate it by taking the log2 of the "Fold change of edge expression weight" but the results are very different. I'll be grateful for any enlightenment on this.

In the second place, I also notice that some partners predicted to interact only in a condition of my data set with respect to the other do not appear as differentially expressed genes when I do a differential gene expression of the interacting partners in the two conditions being tested. Could there be any explanation for this?

Thank you in advance for your help.

asrhou commented 3 years ago

Hi,

Sorry for the late reply. NATMI first calculates the "Fold change of edge expression weight", which is "Edge expression weight in condition 2 / Edge expression weight in condition 1", then "Log2-transformed fold change of edge expression weight" equals "log2(Fold change of edge expression weight)".

I don't quite understand your second question. In NATMI, we only focus on the changes in the edge weights, differentially expressed genes are identified during the calculation.